Report 1: DSIR 16/224: 1/10/1919 to 31/3/1920

ROTARY KILN RESEARCH

In accordance with the programme for the reduction of coal used by rotary kilns, further investigations into the working of kilns and coal dryers have been made during the six months, and drawings for the improvement of various kiln plants have been proceeded with. Details are as follows:-

CONVENTIONAL TERMS

Certain conventions are used in recording the results of experiments on rotary kilns. They are as follows:-

Unit Output - The unit output of a kiln is the clinker produced in cwt per hour, per 1,000 cubic feet of shell capacity, measured inside the lining. Thus a kiln of 5,000 cubic feet capacity which produced 85 cwt of clinker per hour, would have a unit (output) of 85/5 equals 17.0.

Standard Coal - All kiln test results are reduced to percentage (on clinker) of standard coal used. This is dry coal of 7,000 calories, or 12,600 BTU per pound.

Chimney Draught is expressed in "cents". A cent is equal to 1/100 inch of water gauge.

ENQUIRY No.6 - COAL DRYER

APCM plant Swanscombe kilns 1 & 2 had recently been converted into outsized coal dryers, in part prompted by the terrible wet East Midlands duff coal used during the war. This "enquiry" was effectively a commissioning exercise, conducted in November 1919. The project allowed them to tune the dryers to a reasonable performance. Outside of Swanscombe, its usefulness would be restricted to a limited audience of plants that had converted 130' kilns into coal dryers

New type of coal dryer

This investigation relates to a coal dryer 6'6" dia. and 130' long. Powdered coal is used for firing. It is ignited in a furnace tube 2' 3" in dia. and approximately 20 ft. long, which is built into the hood, and projects about 8 feet into the end of the dryer. Cold air for temperature reduction is introduced into the latter half of the tube - the temperature at exit is then 1200° to 1400°F. With the object of reducing dust, only one line of longitudinal lifting plates is fitted to the interior of the shell; they protect 7". The test was undertaken to find out how far the heat of the furnace was applied to the coal by radiation and how far by conduction, also if all portions of the length of the dryer were equally effective.

Gas temperatures along shell

The gas temperatures were measured at five different positions, namely (a) 12" inside the end of the furnace tube, (b) 10' (c) 20' and (d) 40' beyond the end of the furnace tube, and (e) 2" inside the upper end of the dryer shell. In addition to temperature measurements at the centre, the entire cross sections at (a), (b) and (d) were explored by means of a hot air pyrometer. In the end of the furnace tube the temperature varied from 1035° to 1340°F but there was an interval of time between each measurement. At section (b) the temperature varied from 130°F low down near the coal, to 595°F in the highest part of the cross section. At section (d) the temperature similarly varied from 175°F to 506°F.

Coal temperature and moisture along dryer

Inspection doors were fitted to the shell in three positions by means of which coal samples were obtained. Similar samples were taken from the coal entering, and from the coal leaving the dryer. Gas temperature measurements were taken in the same positions. The gas quantities were obtained by measuring the air supply and the rate of coal feed.

Raw coal

The rate of feed was about 8¼ tons per hour.

Method of transferring heat to coal

At the delivery end of the furnace tube about 45% of the total heat of the coal is radiated, and about 50% is in the form of sensible heat in the gases. At the upper end of the dryer about 19% of the total heat of powdered coal is in the form of radiant heat, and this passes to the chimney but this figure includes all losses unaccounted for which may be considerable. About 23% is in the form of sensible heat in the waste gases.

It will be seen therefore that 26% of the heat is transmitted from the flame and heated gases to the coal, by radiation, and 27% is transferred by direct contact from the hot gases to the shell and the coal. It is thought that a dryer tube filled with coal and coal dust would facilitate the transmission to some extent, but the general result is not in accordance with existing theories.

Tests were made on the dryer, on two separate days, for periods of about 8 hours. A heat balance for one test in given below:-

Heat balance for test made 6/11/1919

ENQUIRY No.7 - ROTARY KILN & COAL DRYER

This was BPCM's Wouldham Kiln 9 system. This was Britain's biggest kiln when commissioned in 1912. It was of typical FLS design of the time, with a concentric cooler, which passed most of its hot air to the coal dryer, so that kiln secondary air temperatures were low. The test took place from 21/01/1920 to 02/02/1920. A further "enquiry" - No.13 - took place simultaneously on the coal mill system.

This investigation relates to a rotary kiln and coal drying plant operating on the wet process.

Kiln shell

The general diameter inside the shell plates is 8'11" and the overall length 230'4". There is an enlarged clinkering zone of 10'0" dia. Slurry lifters are provided. The area, reckoned on the leading surface only, of each triangular bracket, is 516 square feet. The capacity of the kiln shell, reckoned inside the lining, is 12,309 cu. ft.

Cooler shell

This is of the usual return tube type, and worked under forced draught. The hot clinker is received on cast iron lining plates which are provided with spirals for the purpose of drawing it in. Air under pressure is supplied to the back of the lining plates for the purpose of keeping them cool. The cooler is large in proportion for the kiln, and is effective.

Hot Air from Cooler

Of the hot air produced by the cooler about 58% is used for coal drying, 26% passes to the kiln, and the remainder is wasted in leakage &c. The hot air traverses the kiln hood by two air ports, and is finally delivered concentric with the coal firing pipe.

Coal Firing Apparatus

There are two screws, each driven by variable speed friction discs. The screws are somewhat remote from the hopper, and it is usual to only have about four to five feet of coal above them, as it is found that a more uniform feed is obtained in this manner.

Dust chamber & chimney-base

The dust chamber is of small dimensions only: the dust caught is about 1/5th of one per cent of the clinker.

Details of test

The kiln and coal dryer were tested for a period of 10 days. The raw coal was weighed in barrow loads of 3 cwt over a platform weigh-bridge.

The clinker was dropped into trucks from the base of the cooler hood, and weighed in lots of about 7 cwt, over a platform weighing machine.

Draught recorders were used to give continuous records:-

(1) in the flue opposite the kiln exit end, (2) in the kiln hood and (3) in the kiln hot air chamber.

Speed of coal feed screws

Recorders were driven from the rim of each friction disc on the two coal feed screws. The maximum speed variation was about 20% of the average speed.

Kiln speed

The kiln has two speeds, one of 0.79 rpm and one of 0.37 rpm. They are obtained by two sets of fast and loose pulleys. A speed recorder, driven from one of the pulleys of the kiln gear, showed that the kiln ran mainly on the fast speed. Hence the feed apparently came down the kiln in a fairly uniform manner.

Kiln test sheet

The more important observations recorded on the test sheet are as follows:-

[Slurry H₂O not given but it was 44.1%]

Temperature Measurements, °F

Air pressures, cents

Draughts, cents

The hot air chamber, and the coal dryer exterior chamber, are built in concrete, and were found to be much less leaky than the chamber usually built in brickwork. The pressure in these chambers is also much higher than usual.

Air quantities in lb per minute

A continuous record was made of the quantity in line (1).

Distribution from hot air chamber in lb per minute

Air supply to kiln lb per minute

The kiln receives hot air from the cooler by the hot air ports and clinker shoot only - see lines (11) and (12) - that is, the hot air supply to the kiln is only 31.4% of the total. The large leakage of cold air into the kiln hood - see line (13) - prevents more hot air being used.

The average temperature of the air entering the kiln is 211°F. The calculated air supply to the kiln, as deduced from the coal burned, and the excess air used, is 1177 lb per minute, which is in good agreement with the value of 1111 lb per minute, as deduced above, from the tilting water gauge measurements.

Details of the calculations are as follows:-

Hence theoretical air supply = 115 × 8.68 × 1.18 = 1177 lb/min

Clinker Rings

The value 3.64 for the ratio SiO₂/R₂O₃ in the slurry, is generally assumed to be too large to admit of the formation of clinker rings. In this kiln they appear from time to time, but usually break away, without causing a stop.

Heat Balance for Cooler

The cooler was of sufficient size, and the radiation losses being small, owing to the design, the apparatus had a high efficiency.

Details are as follows:-

The high efficiency of the cooler is spoiled by the fact that the hot air is largely wasted, partly due to the inefficient coal dryer arrangement.

Radiation from Kiln Shell

The surface temperature of the kiln shell was measured at regular intervals from end to end, and from these measurements the radiation was calculated, using a formula which gives the heat lost in btu per square foot of surface, per hour, for any given value of the shell temperature. The maximum shell temperature was 460°F.

Heat Balance for Kiln

The gain of heat shown in line (13) is usual in the heat balances of long kilns.

The heat balance does not include allowance for the exothermic heat of formation of the clinker minerals from their oxides, which was not known at the time. This contributes to the balancing term (13). In this instance, the exothermic term would be about 1.65% SCC, so the nett heat unaccounted for would be 1.32 - 1.65 = a loss of 0.33% SCC. As implied, the unaccounted heat loss was usually much greater on smaller kilns, mainly because shell losses were consistently underestimated.

General Conclusions drawn from Test

(a) The general performance of the kiln is practically the same as for the shorter kilns of this type, except that the unit output is about 10% low. There is the usual large cold air entry into the kiln hood, due to a fairly large leakage area (9 square feet) and the substantial draught of ¾" water gauge, which is available in the kiln hood.

(b) The relatively high air pressure under which the hot air chamber, and the coal dryer chamber is worked, causes a considerable escape of dust, thus making the burning platform uncomfortable.

(c) The coal dryer arrangement is inefficient, and the coal is not properly dried, There is a considerable loss of coal from the chimney used in connection with the interior of the coal dryer.

(d) The above defects are common to many modern kilns of this type. In other respects this kiln is quite satisfactory, one good feature being the length of time the kiln will run before a stop is necessary for lining renewals.

Saving may be made on the following lines:-

The total coal saving, based on a clinker output of 1400 tons weekly would then be 0.0636 × 1400 = 89 (standard) tons per week.

The improvements suggested would prevent the escape of dust from the hot air chamber, and coal dryer, and also from the coal dryer chimney.

Coal dryer to kiln

The coal dryer test hereafter reported, was made concurrently with the foregoing rotary kiln test, and for the same period of 10 days. The actual running time of the coal dryer was however considerably less than that of the kiln. The dryer is of the usual pattern for plants of this type, and consists essentially of a horizontal steel plate cylindrical shell, which revolves inside a brickwork housing. Hot air from the cooler is used for drying, in two separate streams, one of which is applied to the outside of the shell, and the other passed through the interior. The heat is not properly abstracted from the gases by their contact with the exterior of the shell, and they escape at a relatively high temperature. The temperature of the hot air supply to the interior of the shell is less than it should be, owing to the length of the supply passages. The hot air leaves the interior by a throat casting, which also acts as a feed shoot for raw coal. The escaping air passes up the throat casting, with a relatively high velocity, and meeting the coal stream, a considerable quantity of fine coal and dust is carried out of the dryer, and escapes at the chimney top.

The dimensions inside the shell are:-

(a) Inside diameter 4'6¼"
(b) Length 46'1½"

The internal arrangements comprise 3 rings of inclined lifters, or spirals, at the feed end, the remainder of the shell being occupied by four lines of longitudinal lifters, 10" deep.

The air temperatures entering and leaving the interior, and the exterior of the dryer, were recorded on a quadruple thread recorder, throughout the test.

Test Results

Moisture

Quantity

(7) Coal entering dryer dry tons per hour: 4.30

Size after drying

Temperatures °F

Air quantities lb/min

Speed of Dryer

(17) Revolutions per minute: 3.64

Heat Balance for Test

Suggestions for Improvement

The defects which it is sought to remedy, may be stated as follows:-

(1) The low temperature of the air supplied for drying, 518°F instead of 750°F.

(2) The small abstraction of heat from the gases circulating round the outside of the dryer shell.

(3) Leakage losses, and the heat lost in the long passages leading from the hot air chamber to the interior of the dryer shell.

(4) The escape of coal dust from the chimney leading off the dryer interior.

Taking the points in the order given, improvement may be made as follows on the lines of present practice:-

(1) By reducing the quantity of air passed through the cooler, its temperature will be raised to the required amount. None must then be wasted.

(2) By passing a limited quantity of hot gases, first round the outside of the shell, and afterwards through the interior, a much better abstraction of heat will be obtained. A short well-insulated passage can be arranged, to connect the exterior chamber to the interior of the dryer shell.

(3) The long passages leading from the hot air chamber to the interior of the dryer shell would be put out of use, and sealed off.

(4) The coal firing fan to be arranged to draw as part of its supply, all the air coming from the dryer interior. The coal dust would thus be burned in the kiln. The chimneys leading from the interior, and from the exterior chamber, to be sealed off.

The saving effected is given in the report on the kiln test. In addition the coal will be more completely dried.

ENQUIRY No.8 - 12 ROTARY KILNS

This appears to be APCM's Bevans. The work was not completed, presumably because, after commencing in January 1920, it was pre-empted by the announcement that the whole plant was to be demolished and replaced with new equipment.

An investigation is proceeding in connection with a large plant of 12 short kilns. The kilns being all alike, it is only necessary to investigate one kiln. The arrangements for testing have been drawn out and these comprise a platform over the hopper of the kiln to be tested, which is adapted to carry a weighing machine, for weighing out the powdered coal in 5-cwt lots. The feeding conveyor is suitably inclined, so as to deliver into the top of the weighing hopper.

The clinker, from the battery of each four kilns, is delivered to a shaker conveyor. Arrangements have been drawn out to intercept the clinker from the kiln under test, to elevate it, and to deliver it into a rotary weigher. The weighed clinker to be delivered again to the conveyor.

ENQUIRY No.9 - ROTARY KILN (150 feet)

Probably BPCM's Sundon.

A specification and fifteen prints relating to rotary kiln and coal dryer improvement have been supplied to this works, There is one 150 foot kiln.

ENQUIRY No.10 - ROTARY KILN (200 feet)

This is Johnsons - a BPCM plant. Kiln 4 was added in 1913. It had been the subject of a test ("Enquiry No.4") during 14-20/07/1919. It was proposed to install a new, slightly larger 200' kiln, finally commissioned as Kiln 5 in 1921. This project examined how to add the new kiln using the existing peripheral equipment. Since this was very plant-specific, there was little of value to be applied elsewhere.

There is one 200' kiln in use at this works, and a second 200' kiln is being installed. Drawings are being prepared for the addition of the research improvements to the new Kiln, and for making the existing coal dryer serve for two kilns.

OIL FUEL

They evidently got hit up by this oil burner supplier. There was never any prospect of using oil because of the massive price differential with cheap British coal, which persisted until the 1950s.

A short investigation has been made, with reference to the advantages of oil fuel. They arise from the efficiency with which the combustion can be conducted, and the air supply regulated, always provided that a burner of the proper type is used. It is stated that the supply of air need not be more than 1 lb per lb of fuel in excess of the theoretical. It was stated that certain manufacturers in the north had been discarding coal in favour of oil for boiler and other purposes and the explanation lies in the fact that with coal of very variable quality, and when heat is only required intermittently, the efficiency gained by using a fuel which can be regulated to a nicety, which in constant in quality, and which can be turned on or off at a moment's notice is such compared with the inefficiencies arising from bad coal, unskilled stoking, banking fires, inefficient furnaces etc., that it pays the manufacturers to use oil at two or three times the price of coal.

A sample of fuel oil supplied by the makers of the Hooveler burner, was found to have a calorific value of 8516 calories. A sample of pitch received from the South Metropolitan Gas Co. Ltd. had a calorific value of 8726 calories. A mixture of 50% oil and 50% pitch has been frequently used. £6 per ton is an approximate price for fuel oil at the time of the investigation.

KILN RADIATION EXPERIMENTS

BPCRA always used the word "radiation" inexactly. NPL eventually finished this work, but the definitive scalable model for kiln shell heat loss was done by the US Bureau of Mines.

The National Physical Laboratory, who have this matter in hand, report as follows:-

"The special apparatus constructed consists of a 4' long cylinder, mounted on roller bearings.

Auxiliary gear has been made to drive it at various speeds.

Brush and slip rings are mounted on the apparatus to transmit currents up to 400 amperes to the coil, which will be fitted in the interior of the cylinder.

The various other appliances have been constructed for use in the investigations, and it only remains to wind the heating coil. Considerable difficulty has been met with in finding the materials necessary, and of course, this has led to delay."

GRINDING PLANT RESEARCH

Experimental Tube Mill 18" × 18"

A small tube mill 18" × 18", motor driven, has been installed. It is driven by a 2 BHP variable speed electric motor, and loaded with 300 lb of steel balls of 1" dia. The lining plates are provided with lifter bars, cast on.

Standard sand - 20/30 mesh - is being used as the material for grinding. It is intended to run several tests, using a different weight of sand in each instance, and curves are being obtained which connect the sieve residues with the revolutions of the mill.

ENQUIRY No.11 - TUBE MILL BEARING FRICTION

The ultimate objective of grinding plant research was to identify the energy expended by tube mills on actual grinding. It was therefore first necessary to derive a model for the power absorbed in just turning the loaded mill, without any grinding action. At this early stage of development, mill trunnion bearings were still primitive, and presumably inefficient.

The object of this test was to determine the horse power in a tube mill due to bearing friction. The method adopted was to build in the tube a concentric brickwork load, equal in weight to the grinding media, and the material under treatment. The horse power required to rotate the mill under these conditions should be that due to friction alone. The dimensions of the mill are as follows:-

(1) Internal dimensions of tube: dia 72½", length 15'6"
(2) Weight of brickwork in tube: 32.51 ton
(3) Main bearing, feed end: dia 20 15/16", length 18¾"
(4) Main bearing, delivery end: dia 20¾", length 20½"

It was considered that a test of a mill of this type would be interesting, on account of the relatively large diameter of the main bearings, and the absence of a tyre and friction rollers. The mill is driven by a 260 HP motor, through an intermediate countershaft. The main bearings are lubricated by a block of grease, resting on the top of the journal. It was arranged to drive the mill at various speeds, so as to obtain a comparison of the frictional resistance with the speed. The mill speed was therefore varied between 24 and 29 rpm. The frictional resistance was taken both with the tube mill empty, and also when containing 5, 10 and 15 tons of brickwork. The horse powers absorbed by the countershaft, pinion shaft, and rope drives were also noted.

Description of Tests

Approximately 16 sets of observations were taken during each of the BHP tests, the motor speed being varied through a considerable range in each case. The test results were plotted as BHP against motor speed, but the HP exerted was not in any way proportional to the speed. An average value of the BHP for a mill speed of 25 rpm is however obtained in this manner.

At the conclusion of the tests above described, the ropes were removed from the motor pulley, and a series of no-load tests made on the motor. The speed was varied by the shunt resistance between 390 and 506 rpm.

The no-load tests were taken, so that the actual net horsepower used by the tube mill could be obtained.

A summary of the results, averaged out for a mill speed of 25 rpm is as follows:-

Table of results: mill speed 25 rpm: mill bearings velocity of rubbing 2.27 ft/ton (sic)

• column (1) = mill iron parts only
• column (2) = mill iron parts + 1/3 brickwork load
• column (3) = mill iron parts + 2/3 brickwork load
• column (4) = mill iron parts + full brickwork load

The "velocity of rubbing" = average circumference of bearings × rotation rate
= π × (20.9375 + 20.75) / 2 × 25 / 60 / 12 = 2.2737 ft/s. The fact that the frictional force diminished both at high speed and at high load indicates that multiple slippage mechanisms are at work, so the results probably could not be made generally applicable.

THEORY OF GRINDING ACTION

This line of enquiry led eventually to the surface area model for grindability. Early progress was severely hampered by the lack of precise methods of particle size analysis below the 90-mμ size of the 180-mesh sieve, which was the finest in use at the time. From well back in the 19th century it had been recognised that cement performance was mostly controlled by the amount of "flour" - say, material below 20 mμ - but this could only be assessed by crude, qualitative methods. Specific surface is almost entirely governed by the amount of theses fines, and until the air-permeability test was developed in the late 1930s, useful data could not be obtained.

Mr Chas. E. Blyth has put before the Research Council two new theories for estimating the work done in crushing and grinding. They are the "New Surface" Theory, and Stadler's Theory.

Charles Edward Blyth was a director of Charles Nelson & Co and manager of Stockton plant.

New Surface Theory is the classification of the various sizes of particles into grades in accord with the surface exposed, based upon the theory that "The work done in crushing and grinding is in proportion to the new surface exposed".

Let E = the unit of work done, i.e., the energy required to overcome the resistance of the cohesion of the molecules over an area 1 square inch

S = the new surface produced

N = the number of pieces or cubes to the lineal inch

Taking for the purpose of illustration a volume of 1 cubic inch (which might be composed of any sized particles, but for the sake of this argument it is conceded to be 1 inch cube). By the expenditure of unit energy (E) the volume is divided producing two [square] inches of new surface, two further units of energy utilised in the most economical way will reduce the original volume to its next analogous state (cubic) the volume of each being ⅛ of the original cube. It is shown therefore that:

E = S/2 = 3 (N-1)

or S = 2E = 6 (N-1)

or N = E/3+1= S/6+1

These formulae give for each sieve used, a grading value, considering the mesh aperture as the size of the piece, which is just unable to pass through the screen.

The sieve sizes are somewhat off, suggesting that they were using non-standard sieves. The standard BS12 sieves of the time had apertures:
20# 0.03360": 30# 0.02253": 50# 0.01240": 76# 0.00876": 100# 0.00600": 120# 0.00493": 180# 0.00356".
The psd work also shows that the sieves were not correctly calibrated.

For the value of the fines passing the 180# sieve, since 3 units of energy are required to produce an analogous range so that the whole volume will pass to the grade below, so expenditure of 1.5 units will be required to deal with half the volume, but as the volume of the finer particles is now equal to the coarser fragments further energy expended will be absorbed by both fine and coarse material in proportion to their respective volumes. Working on these lines a table may be calculated as follows:-

These figures will give the points of a curve, from which a multiplier may be found, for any percentage passing a screen, viz. 50 per cent passing 180 mesh, value 854 × 1.5 = 1281 units.

I have spent long hours trying to find some fragments of logic in the above, but it is all complete nonsense. The success of the "new surface" theory depends critically upon being able to assign a surface to the fines, and this is an attempt to persuade the reader that the problem had been solved. All it actually does is to assign a size of about 30 μm to the material passing 180# (90 μm), diminishing to somewhat below 20 μm when more than 98% passes. A good test of this approach is to ask the question - what would be the size if 100% passed 180#? Of course the correct answer is - somewhere between 90 μm and zero. But within that range, the value is completely indeterminate.

As to how far the new surface produced is proportional the energy expended, was tested by means of a small tube mill 20" dia by 12" long. Standard sand through 20 mesh, and retained on 30 mesh, was used as the material to be ground.

The mill was loaded with 100 lb of Holpebs and 35 lb of sand, and the grinding result was examined at the end of each 500 revolutions.

Tests

On completing 500 revolutions the mill was stopped and 10 oz of the material was withdrawn and replaced with 10 oz of fresh standard sand, the mill was then sent off on the next trip of 500 revolutions.

The sample retained was tested for residue on the following sieves - 30, 50, 76, 100, 120 and 180 mesh, 1 oz of fines capable of passing 180 mesh sieve were reserved for a further test by "sedimentation". The residue on the sieves together with the rest of the sample was made up to 10 oz with fresh sand and used to replace the sample taken from the mill in the subsequent run.

After 8,500 revs. the mill was emptied, and the whole of the ground material was thoroughly turned over, and then sampled in the usual way.

Grading Analysis

A table followed showing the seven sieve size percent gradings every 500 revs from 500 to 8500 revs, with calculated surface area for each. The table is far too large to present in html form, besides which it is not particularly informative. It is available as an Excel file on application.

Examination of the "fines" was carried out on the lines laid out by A D Hall as described in the Journal of the Chemical Society, LXXXV (1904) p 950.

Alfred Daniel Hall worked at the Rothamsted Experimental Station, and his method was designed for soil analysis. It consisted simply of agitating the sample in water in a beaker, and decanting the un-settled supernatant layer after fixed times of settlement, these times being calculated to yield standard particle size cuts in accordance with Stokes' spherical particle model. Acknowledged as imprecise but taking several weeks to perform, it was fine for Rothamsted's century-long experiments, but for cement it yielded questionable results, as demonstrated by their eccentric Rosin/Rammler plots.

Taking the size of the particles as derived from this test and using the formula E = 3(N-1) gives the following energy values:-

Another table followed - again too big to reproduce, but available in the Excel file.

Taking the mechanical value given by the sedimentation process and dividing each result by the value of the 180 mesh sieve we get a figure to compare with the multiplier taken from the graph; they are fairly comparative with the exception of the result of the 5th run, in which test there appears to be quite an undue amount of ultra fine material, the mechanical value of which is some 1140 units of energy higher than the previous test, and is probably due to some accident in sampling.

Comparison

This New Surface Theory is being investigated and developed.

The table gives a comforting confirmation of the graphical method of divining the fineness of the -180# material. However, the numbers in the table don't appear to correspond very well with those actually found. Here are the real numbers. Some of the errors are just mis-typing. But many are not. I make no comment.

ENQUIRY No.13 (given as 11) - COAL GRINDING PLANT

This was conducted at BPCM's Wouldham concurrently with the kiln/coal dryer examination in Enquiry No.7 during 21/01/1920 to 02/02/1920, the final report was delayed due to tedious data processing requirements, and the final results were given in Report No.2

This coal grinding plant consist of a Kominor and a tube mill. It was tested for 10 days, from 21st January to 2nd February, and concurrently with the kiln. The coal ground, reckoned dry, was 4.32 tons per hour, to a residue of 13 per cent on 180#, with an expenditure of 175.6 BHP. There has been some delay in presenting a full report of the test, the method of working out the new surface produced, using Hall's method, necessitating considerable time and care.

RAW MATERIAL RESEARCH

The initial approach seems to have been to address the question of how rawmix chemistry should be set for rotary kilns - a question still unresolved at the time.

The work done during the past half year has mainly consisted in the study of the bibliography of the subject, and in the preparation of the special apparatus required. It has already been fairly well established by various authorities, notably the US Bureau of Standards, that Portland cement consists essentially of:-

• (a) tricalcium silicate 3CaO.SiO₂
• (b) calcium orthosilicate 2CaO.SiO₂
• (c) tricalcium aluminate 3CaO.Al₂O₃

in certain somewhat restricted proportions.

Of these, tricalcium silicate has itself all the properties of good Portland cement, e.g. - it is readily hydrated; is sound; and rapidly attains high tensile and crushing strength.

Calcium orthosilicate hydrates very slowly, and has no strength, to speak of, up to 28 days or more, but thereafter develops considerable and increasing strength.

Tricalcium aluminate hydrates very rapidly, imparting quick initial setting to cement. It has little or no compressive or adhesive strength, and does not harden under water.

It is suggested that an experimental kiln, capable of producing sufficient clinker for tensile or crushing tests, soundness and setting time tests &c. be provided, and experiments made on the following lines:-

(1) Experimental burnings of normal slurry, or dry mixtures, for comparison with cement produced commercially, from the same materials, and in the same proportions.

(2) Effect of variations in the proportions of the acid and basic materials:-

(a) Increase of lime, by steps of, say 2%
(b) Increase of silica in a similar manner.

(3) Substitution of other fluxes for alumina

(a) Substitution of ferric oxide
(b) Other possible substitutes

(4) Substitution of magnesia for lime

Substitution of baryta for lime

(5) Effect of variations in the rate of cooling clinker. Effect of variation in the temperature and time of burning.

(6) Endeavour to find a 'barrier" or" catalyser" the addition of which, in small proportions, would facilitate combination, and thus reduce the temperature necessary in the clinkering zone. (example:- the addition of 1% of Arsenious Oxide in glass manufacture)

(7) To repeat on the large scale any of the above which promise commercial success.

It might be considered later advisable to carry out microscopical examination of clinker sections, and the measurement of optical properties, to ascertain the nature of the crystalline products: but this is not deemed necessary at present.